Chapter 1

INTRODUCTION
1.1 Robotics

Robotics is a relatively young eld of modern technology that crosses traditional engineering boundaries. Understanding the complexity of robots and their applications requires knowledge of electrical engineering, mechanical engineering, systems and industrial engineering, computer science, economics, and mathematics. New disciplines of engineering, such as manufacturing engineering, applications engineering, and knowledge engineering have emerged to deal with the complexity of the eld of robotics and factory automation. This book is concerned with fundamentals of robotics, including kinematics, dynamics, motion planning, computer vision, and control. Our goal is to provide a complete introduction to the most important concepts in these subjects as applied to industrial robot manipulators. The science of robotics has grown tremendously over the past twenty years, fueled by rapid advances in computer and sensor technology as well as theoretical advances in control and computer vision. In addition to the topics listed above, robotics encompasses several areas not covered in this text such as locomotion, including wheeled and legged robots, ying and swimming robots, grasping, articial intelligence, computer architectures, programming languages, and computer-aided design. A complete treatment of the discipline of robotics would require several volumes. Nevertheless, at the present time, the vast majority of robot applications deal with industrial robot arms operating in structured factory environments so that a rst introduction to the subject of robotics must include a rigorous treatment of the topics in this text. 3

CHAPTER 1. INTRODUCTION

1.2

History of Robotics

The term robot was rst introduced into our vocabulary by the Czech playwright Karel Capek in his 1920 play Rossums Universal Robots, the word robota being the Czech word for work. Since then the term has been applied to a great variety of mechanical devices, such as teleoperators, underwater vehicles, autonomous land rovers, etc. Virtually anything that operates with some degree of autonomy, usually under computer control, has at some point been called a robot. In this text the term robot will mean a computer controlled industrial manipulator of the type shown in Figure 1.1. This type of robot is essentially a mechanical arm operating under computer control. Such devices, though far from the robots of science ction, are nevertheless extremely complex electro-mechanical systems whose analytical description requires advanced methods, and which present many challenging and interesting research problems.

Figure 1.1: The ABB IRB6600 Robot. Photo courtesy of ABB

1.2. HISTORY OF ROBOTICS

An ocial denition of such a robot comes from the Robot Institute of America (RIA): A robot is a reprogrammable multifunctional manipulator designed to move material, parts, tools, or specialized devices through variable programmed motions for the performance of a variety of tasks. The key element in the above denition is the reprogrammability of robots. It is the computer brain that gives the robot its utility and adaptability. The so-called robotics revolution is, in fact, part of the larger computer revolution. Even this restricted version of a robot has several features that make it attractive in an industrial environment. Among the advantages often cited in favor of the introduction of robots are decreased labor costs, increased precision and productivity, increased exibility compared with specialized machines, and more humane working conditions as dull, repetitive, or hazardous jobs are performed by robots. The robot, as we have dened it, was born out of the marriage of two earlier technologies: that of teleoperators and numerically controlled milling machines. Teleoperators, or master-slave devices, were developed during the second world war to handle radioactive materials. Computer numerical control (CNC) was developed because of the high precision required in the machining of certain items, such as components of high performance aircraft. The rst robots essentially combined the mechanical linkages of the teleoperator with the autonomy and programmability of CNC machines. Several milestones on the road to present day robot technology are listed below.

Milestones in the History of Robotics

1947 the rst servoed electric powered teleoperator is developed 1948 a teleoperator is developed incorporating force feedback 1949 research on numerically controlled milling machine is initiated 1954 George Devol designs the rst programmable robot 1956 Joseph Engelberger, a Columbia University physics student, buys the rights to Devols robot and founds the Unimation Company 1961 the rst Unimate robot is installed in a Trenton, New Jersey plant of General Motors to tend a die casting machine 1961 the rst robot incorporating force feedback is developed 1963 the rst robot vision system is developed 1971 the Stanford Arm is developed at Stanford University

CHAPTER 1. INTRODUCTION

1973 the rst robot programming language (WAVE) is developed at Stanford 1974 Cincinnati Milacron introduced the T 3 robot with computer control 1975 Unimation Inc. registers its rst nancial prot 1976 the Remote Center Compliance (RCC) device for part insertion in assembly is developed at Draper Labs in Boston 1976 Robot arms are used on the Viking I and II space probes and land on Mars 1978 Unimation introduces the PUMA robot, based on designs from a General Motors study 1979 the SCARA robot design is introduced in Japan 1981 the rst direct-drive robot is developed at Carnegie-Mellon University 1982 Fanuc of Japan and General Motors form GM Fanuc to market robots in North America 1983 Adept Technology is founded and successfully markets the direct drive robot 1986 the underwater robot, Jason, of the Woods Hole Oceanographic Institute, explores the wreck of the Titanic, found a year earlier by Dr. Robert Barnard. 1988 St aubli Group purchases Unimation from Westinghouse 1988 the IEEE Robotics and Automation Society is formed 1993 the experimental robot, ROTEX, of the German Aerospace Agency (DLR) was own aboard the space shuttle Columbia and performed a variety of tasks under both teleoperated and sensor-based oine programmed modes 1996 Honda unveils its Humanoid robot; a project begun in secret in 1986 1997 the rst robot soccer competition, RoboCup-97, is held in Nagoya, Japan and draws 40 teams from around the world 1997 the Sojourner mobile robot travels to Mars aboard NASAs Mars PathFinder mission 2001 Sony begins to mass produce the rst household robot, a robot dog named Aibo

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

2001 the Space Station Remote Manipulation System (SSRMS) is launched in space on board the space shuttle Endeavor to facilitate continued construction of the space station 2001 the rst telesurgery is performed when surgeons in New York performed a laparoscopic gall bladder removal on a woman in Strasbourg, France 2001 robots are used to search for victims at the World Trade Center site after the September 11th tragedy 2002 Hondas Humanoid Robot ASIMO rings the opening bell at the New York Stock Exchange on February 15th The rst successful applications of robot manipulators generally involved some sort of material transfer, such as injection molding or stamping where the robot merely attended a press to unload and either transfer or stack the nished part. These rst robots were capable of being programmed to execute a sequence of movements, such as moving to a location A, closing a gripper, moving to a location B, etc., but had no external sensor capability. More complex applications, such as welding, grinding, deburring, and assembly require not only more complex motion but also some form of external sensing such as vision, tactile, or force-sensing, due to the increased interaction of the robot with its environment. It should be pointed out that the important applications of robots are by no means limited to those industrial jobs where the robot is directly replacing a human worker. There are many other applications of robotics in areas where the use of humans is impractical or undesirable. Among these are undersea and planetary exploration, satellite retrieval and repair, the defusing of explosive devices, and work in radioactive environments. Finally, prostheses, such as articial limbs, are themselves robotic devices requiring methods of analysis and design similar to those of industrial manipulators.

1.3
1.3.1

Components and Structure of Robots

Symbolic Representation of Robots

Robot Manipulators are composed of links connected by joints into a kinematic chain. Joints are typically rotary (revolute) or linear (prismatic). A revolute joint is like a hinge and allows relative rotation between two links. A prismatic joint allows a linear relative motion between two links. We

CHAPTER 1. INTRODUCTION

use the convention (R) for representing revolute joints and (P ) for prismatic joints and draw them as shown in Figure 1.2.

Revolute

Prismatic

2D

3D

Figure 1.2: Symbolic representation of robot joints.

Each joint represents the interconnection between two links, say i and i+1 . We denote the axis of rotation of a revolute joint, or the axis along which a prismatic joint slides by zi if the joint is the interconnection of links i and i + 1. The joint variables, denoted by i for a revolute joint and di for the prismatic joint, represent the relative displacement between adjacent links. We will make this precise in Chapter 3.

1.3.2

Degrees of Freedom and Workspace

The number of joints determines the degrees-of-freedom (DOF) of the manipulator. Typically, a manipulator should possess at least six independent DOF: three for positioning and three for orientation. With fewer than six DOF the arm cannot reach every point in its work environment with arbitrary orientation. Certain applications such as reaching around or behind obstacles require more than six DOF. The diculty of controlling a manipulator increases rapidly with the number of links. A manipulator having more than six links is referred to as a kinematically redundant manipulator.

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

The workspace of a manipulator is the total volume swept out by the end-eector as the manipulator executes all possible motions. The workspace is constrained by the geometry of the manipulator as well as mechanical constraints on the joints. For example, a revolute joint may be limited to less than a full 360 of motion. The workspace is often broken down into a reachable workspace and a dextrous workspace. The reachable workspace is the entire set of points reachable by the manipulator, whereas the dextrous workspace consists of those points that the manipulator can reach with an arbitrary orientation of the end-eector. Obviously the dextrous workspace is a subset of the reachable workspace. The workspaces of several robots are shown later in this chapter.

1.3.3

Classication of Robots

Robot manipulators can be classied by several criteria, such as their power source, or way in which the joints are actuated, their geometry, or kinematic structure, their intended application area, or their method of control. Such classication is useful primarily in order to determine which robot is right for a given task. For example, an hydraulic robot would not be suitable for food handling or clean room applications, whereas a SCARA robot would not be suitable for work in a foundry. We explain this in more detail below. (i) Power Source Typically, robots are either electrically, hydraulically, or pneumatically powered. Hydraulic actuators are unrivaled in their speed of response and torque producing capability. Therefore hydraulic robots are used primarily for lifting heavy loads. The drawbacks of hydraulic robots are that they tend to leak hydraulic uid, require much more peripheral equipment, such as pumps, which also requires more maintenance, and they are noisy. Robots driven by DC- or AC-servo motors are increasingly popular since they are cheaper, cleaner and quieter. Pneumatic robots are inexpensive and simple but cannot be controlled precisely. As a result, pneumatic robots are limited in their range of applications and popularity. (ii) Application Area The largest projected area of future application of robots is in assembly. Therefore, robots are often classied by application into assembly and nonassembly robots. Assembly robots tend to be small, electrically driven and

10

CHAPTER 1. INTRODUCTION

either revolute or SCARA (described below) in design. The main nonassembly application areas to date have been in welding, spray painting, material handling, and machine loading and unloading. (iii) Method of Control Robots are classied by control method into servo and non-servo robots. The earliest robots were non-servo robots. These robots are essentially openloop devices whose movement is limited to predetermined mechanical stops, and they are useful primarily for materials transfer. In fact, according to the denition given previously, xed stop robots hardly qualify as robots. Servo robots use closed-loop computer control to determine their motion and are thus capable of being truly multifunctional, reprogrammable devices. Servo controlled robots are further classied according to the method that the controller uses to guide the end-eector. The simplest type of robot in this class is the point-to-point robot. A point-to-point robot can be taught a discrete set of points but there is no control on the path of the end-eector in between taught points. Such robots are usually taught a series of points with a teach pendant. The points are then stored and played back. Point-to-point robots are severely limited in their range of applications. In continuous path robots, on the other hand, the entire path of the end-eector can be controlled. For example, the robot endeector can be taught to follow a straight line between two points or even to follow a contour such as a welding seam. In addition, the velocity and/or acceleration of the end-eector can often be controlled. These are the most advanced robots and require the most sophisticated computer controllers and software development. (iv) Geometry Most industrial manipulators at the present time have six or fewer degreesof-freedom. These manipulators are usually classied kinematically on the basis of the rst three joints of the arm, with the wrist being described separately. The majority of these manipulators fall into one of ve geometric types: articulate (RRR), spherical (RRP), SCARA (RRP), cylindrical (RPP), or cartesian (PPP). We discuss each of these in detail below. Each of these ve congurations are serial link robots. A sixth and fundamentally distinct class of manipulators is the so-called parallel robot. In a parallel conguration the links are arranged in a closed rather than open kinematic chain. We

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

11

include a discussion of the parallel robot for completeness as parallel robots are becoming increasingly common.

1.3.4

Common Kinematic Arrangements

Articulated Conguration (RRR) The articulated manipulator is also called a revolute, or anthropomorphic manipulator. The ABB IRB1400 articulated arm is shown in Figure 1.3. A common revolute joint design is the parallelogram linkage

Figure 1.3: The ABB IRB1400 Robot. Photo courtesy of ABB such as the Motoman SK16, shown in Figure 1.4. In both of these arrangements joint axis z2 is parallel to z1 and both z1 and z2 are perpendicular to z0 . The structure and terminology associated with the elbow manipulator are shown in Figure 1.5. Its workspace is shown in Figure 1.6. The revolute conguration provides for relatively large freedom of movement in a compact space. The parallelogram linkage, although less dextrous typically than the elbow manipulator conguration, nevertheless has several advantages that make it an attractive and popular design. The most notable feature of the

12

CHAPTER 1. INTRODUCTION

Figure 1.4: The Motoman SK16 manipulator.

parallelogram linkage conguration is that the actuator for joint 3 is located on link 1. Since the weight of the motor is born by link 1, links 2 and 3 can be made more lightweight and the motors themselves can be less powerful. Also the dynamics of the parallelogram manipulator are simpler than those of the elbow manipulator, thus making it easier to control.

Spherical Conguration (RRP) By replacing the third or elbow joint in the revolute conguration by a prismatic joint one obtains the spherical conguration shown in Figure 1.7. The term spherical conguration derives from the fact that the spherical coordinates dening the position of the end-eector with respect to a frame whose origin lies at the intersection of the axes z1 and z2 are the same as the rst three joint variables. Figure 1.8 shows the Stanford Arm, one of the most well-known spherical robots. The workspace of a spherical manipulator is shown in Figure 1.9.

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

13

z0 2 Shoulder Forearm Elbow 1 Body Base

Figure 1.5: Structure of the elbow manipulator.

z1

z2

SCARA Conguration (RRP) The so-called SCARA (for Selective Compliant Articulated Robot for Assembly) shown in Figure 1.10 is a popular conguration, which, as its name suggests, is tailored for assembly operations. Although the SCARA has an RRP structure, it is quite dierent from the spherical conguration in both appearance and in its range of applications. Unlike the spherical design, which has z0 , z1 , z2 mutually perpendicular, the SCARA has z0 , z1 , z2 parallel. Figure 1.11 shows the Epson E2L653S, a manipulator of this type. The SCARA manipulator workspace is shown in Figure 1.12.

Cylindrical Conguration (RPP) The cylindrical conguration is shown in Figure 1.13. The rst joint is revolute and produces a rotation about the base, while the second and third joints are prismatic. As the name suggests, the joint variables are the cylindrical coordinates of the end-eector with respect to the base. A cylindrical robot, the Seiko RT3300, is shown in Figure 1.14, with its workspace shown in Figure 1.15.

14

CHAPTER 1. INTRODUCTION

Top

Side

Figure 1.6: Workspace of the elbow manipulator. Cartesian conguration (PPP) A manipulator whose rst three joints are prismatic is known as a cartesian manipulator, shown in Figure 1.16. For the Cartesian manipulator the joint variables are the Cartesian coordinates of the end-eector with respect to the base. As might be expected the kinematic description of this manipulator is the simplest of all congurations. Cartesian congurations are useful for table-top assembly applications and, as gantry robots, for transfer of material or cargo. An example of a cartesian robot, from Epson-Seiko, is shown in Figure 1.17. The workspace of a Cartesian manipulator is shown in Figure 1.18. Parallel Manipulator A parallel manipulator is one in which the links form a closed chain. More specically, a parallel manipulator has two or more independent kinematic chains connecting the base to the end-eector. Figure 1.19 shows the ABB IRB 940 Tricept robot, which has a parallel conguration. The closed chain kinematics of parallel robots can result in greater structural rigidity, and hence greater accuracy, than open chain robots. The kinematic description of parallel robots fundamentally dierent from that of serial link robots and

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

15

z0 2

z1 d3 z2

Figure 1.7: The spherical manipulator conguration. therefore requires dierent methods of analysis.

1.3.5

Robotic Systems

A robot manipulator should be viewed as more than just a series of mechanical linkages. The mechanical arm is just one component to an overall Robotic System, shown in Figure 1.20, which consists of the arm, external power source, end-of-arm tooling, external and internal sensors, computer interface, and control computer. Even the programmed software should be considered as an integral part of the overall system, since the manner in which the robot is programmed and controlled can have a major impact on its performance and subsequent range of applications.

1.3.6

Accuracy and Repeatability

The accuracy of a manipulator is a measure of how close the manipulator can come to a given point within its workspace. Repeatability is a measure of how close a manipulator can return to a previously taught point. Most present day manipulators are highly repeatable but not very accurate. The primary method of sensing positioning errors in most cases is with position

16

CHAPTER 1. INTRODUCTION

Figure 1.8: The Stanford Arm. Photo courtesy of the Coordinated Science Lab, University of Illinois at Urbana-Champaign. encoders located at the joints, either on the shaft of the motor that actuates the joint or on the joint itself. There is typically no direct measurement of the end-eector position and orientation. One must rely on the assumed geometry of the manipulator and its rigidity to infer (i.e., to calculate) the end-eector position from the measured joint angles. Accuracy is aected therefore by computational errors, machining accuracy in the construction of the manipulator, exibility eects such as the bending of the links under gravitational and other loads, gear backlash, and a host of other static and dynamic eects. It is primarily for this reason that robots are designed with extremely high rigidity. Without high rigidity, accuracy can only be improved by some sort of direct sensing of the end-eector position, such as with vision. Once a point is taught to the manipulator, however, say with a teach pendant, the above eects are taken into account and the proper encoder values necessary to return to the given point are stored by the controlling computer. Repeatability therefore is aected primarily by the controller resolution. Controller resolution means the smallest increment of mo-

1.3. COMPONENTS AND STRUCTURE OF ROBOTS

17

Figure 1.9: Workspace of the spherical manipulator.

tion that the controller can sense. The resolution is computed as the total distance traveled by the tip divided by 2n , where n is the number of bits of encoder accuracy. In this context, linear axes, that is, prismatic joints, typically have higher resolution than revolute joints, since the straight line distance traversed by the tip of a linear axis between two points is less than the corresponding arclength traced by the tip of a rotational link. In addition, as we will see in later chapters, rotational axes usually result in a large amount of kinematic and dynamic coupling among the links with a resultant accumulation of errors and a more dicult control problem. One may wonder then what the advantages of revolute joints are in manipulator design. The answer lies primarily in the increased dexterity and compactness of revolute joint designs. For example, Figure 1.21 shows that for the same range of motion, a rotational link can be made much smaller than a link with linear motion. Thus manipulators made from revolute joints occupy a smaller working volume than manipulators with linear axes. This increases the ability of the manipulator to work in the same space with other robots, machines, and people. At the same time revolute joint manipulators are better able to maneuver around obstacles and have a wider range of possible applications.

18

CHAPTER 1. INTRODUCTION

z1 2

z2

d3 z0 1

Figure 1.10: The SCARA (Selective Compliant Articulated Robot for Assembly).

1.3.7

Wrists and End-Eectors

The wrist of a manipulator refers to the joints in the kinematic chain between the arm and hand. The wrist joints are nearly always all revolute. It is increasingly common to design manipulators with spherical wrists, by which we mean wrists whose three joint axes intersect at a common point. The spherical wrist is represented symbolically in Figure 1.22. The spherical wrist greatly simplies the kinematic analysis, eectively allowing one to decouple the positioning and orientation of an object to as great an extent as possible. Typically therefore, the manipulator will possess three positional degrees-of-freedom, which are produced by three or more joints in the arm. The number of orientational degrees-of-freedom will then depend on the degrees-of-freedom of the wrist. It is common to nd wrists having one, two, or three degrees-of-freedom depending of the application. For example, the SCARA robot shown in Figure 1.11 has four degrees-offreedom: three for the arm, and one for the wrist, which has only a roll about the nal z -axis. It has been said that a robot is only as good as its hand or end-eector. The arm and wrist assemblies of a robot are used primarily for positioning the end-eector and any tool it may carry. It is the end-eector or tool that actually performs the work. The simplest type of end-eectors are grippers, such as shown in Figures 1.23 and 1.24, which usually are capable of only two actions, opening and closing. While this is adequate for materials transfer, some parts handling, or gripping simple tools, it is not adequate for other tasks such as welding, assembly, grinding, etc. A great deal of

1.4. OUTLINE OF THE TEXT

19

Figure 1.11: The Epson E2L653S SCARA Robot. Photo Courtesy of Epson. research is therefore being devoted to the design of special purpose endeectors as well as tools that can be rapidly changed as the task dictates. There is also much research being devoted to the development of anthropomorphic hands. Such hands are being developed both for prosthetic use and for use in manufacturing. Since we are concerned with the analysis and control of the manipulator itself and not in the particular application or end-eector, we will not discuss end-eector design or the study of grasping and manipulation.

1.4

Outline of the Text

A typical application involving an industrial manipulator is shown in Figure 1.25. The manipulator is shown with a grinding tool which it must use to remove a certain amount of metal from a surface. In the present text we are concerned with the following question: What are the basic issues to be resolved and what must we learn in order to be able to program a robot to perform tasks such as the above? The ability to answer this question for a full six degree-of-freedom manipulator represents the goal of the present text. The answer itself is too complicated to be presented at this point. We can, however, use the simple two-link planar mechanism to illustrate some of the major issues involved and to preview the topics covered in this text. Suppose we wish to move the manipulator from its home position to

20

CHAPTER 1. INTRODUCTION

Figure 1.12: Workspace of the SCARA manipulator. position A, from which point the robot is to follow the contour of the surface S to the point B , at constant velocity, while maintaining a prescribed force F normal to the surface. In so doing the robot will cut or grind the surface according to a predetermined specication.

Problem 1: Forward Kinematics

The rst problem encountered is to describe both the position of the tool and the locations A and B (and most likely the entire surface S ) with respect to a common coordinate system. In Chapter 2 we give some background on representations of coordinate systems and transformations among various coordinate systems. Typically, the manipulator will be able to sense its own position in some manner using internal sensors (position encoders) located at joints 1 and 2, which can measure directly the joint angles 1 and 2 . We also need therefore to express the positions A and B in terms of these joint angles. This leads to the forward kinematics problem studied in Chapter 3, which is to determine the position and orientation of the end-eector or tool in terms of the joint variables. It is customary to establish a xed coordinate system, called the world or base frame to which all objects including the manipulator are referenced.

1.4. OUTLINE OF THE TEXT

21

d3 z2

z1 d2 z0 1

Figure 1.13: The cylindrical manipulator conguration. In this case we establish the base coordinate frame o0 x0 y0 at the base of the robot, as shown in Figure 1.26. The coordinates (x, y ) of the tool are expressed in this coordinate frame as x = x2 = 1 cos 1 + 2 cos(1 + 2 ) y = y2 = 1 sin 1 + 2 sin(1 + 2 ), (1.1) (1.2)

in which 1 and 2 are the lengths of the two links, respectively. Also the orientation of the tool frame relative to the base frame is given by the direction cosines of the x2 and y2 axes relative to the x0 and y0 axes, that is, x2 x0 = cos(1 + 2 ); y2 x0 = x2 y0 = sin(1 + 2 ) (1.3)

sin(1 + 2 );

y2 y0 = sin(1 + 2 )

which we may combine into an orientation matrix x2 x0 y2 x0 x2 y0 y2 y0 = cos(1 + 2 ) sin(1 + 2 ) sin(1 + 2 ) cos(1 + 2 ) . (1.4)

22

CHAPTER 1. INTRODUCTION

Figure 1.14: The Seiko RT3300 Robot. Photo courtesy of Seiko. These equations (1.1-1.4) are called the forward kinematic equations. For a six degree-of-freedom robot these equations are quite complex and cannot be written down as easily as for the two-link manipulator. The general procedure that we discuss in Chapter 3 establishes coordinate frames at each joint and allows one to transform systematically among these frames using matrix transformations. The procedure that we use is referred to as the Denavit-Hartenberg convention. We then use homogeneous coordinates and homogeneous transformations to simplify the transformation among coordinate frames.

Problem 2: Inverse Kinematics

Now, given the joint angles 1 , 2 we can determine the end-eector coordinates x and y . In order to command the robot to move to location B we need the inverse; that is, we need the joint variables 1 , 2 in terms of the x and y coordinates of B . This is the problem of Inverse Kinematics. In other words, given x and y in the forward kinematic equations (1.1-1.2), we wish to solve for the joint angles. Since the forward kinematic equations

1.4. OUTLINE OF THE TEXT

23

Figure 1.15: Workspace of the cylindrical manipulator. are nonlinear, a solution may not be easy to nd nor is there a unique solution in general. We can see, for example, in the case of a two-link planar mechanism that there may be no solution, if the given (x, y ) coordinates are out of reach of the manipulator. If the given (x, y ) coordinates are within the manipulators reach there may be two solutions as shown in Figure 1.27, the so-called elbow up and elbow down congurations, or there may be exactly one solution if the manipulator must be fully extended to reach the point. There may even be an innite number of solutions in some cases (Problem 1.25). Consider the diagram of Figure 1.28. Using the Law of Cosines we see that the angle 2 is given by cos 2 =
2 2 x2 + y 2 1 2 := D. 21 2

(1.5)

We could now determine 2 as 2 = cos1 (D). (1.6)

24

CHAPTER 1. INTRODUCTION

d2 z1 d1 z0 d3

z2

Figure 1.16: The cartesian manipulator conguration. However, a better way to nd 2 is to notice that if cos(2 ) is given by (1.5) then sin(2 ) is given as sin(2 ) = 1 D2 and, hence, 2 can be found by 2 = tan
1

(1.7)

1 D2 . D

(1.8)

The advantage of this latter approach is that both the elbow-up and elbow-down solutions are recovered by choosing the positive and negative signs in (1.8), respectively. It is left as an exercise (Problem 1.19) to show that 1 is now given as 1 = tan1 (y/x) tan1 2 sin 2 . 1 + 2 cos 2 (1.9)

Notice that the angle 1 , depends on 2 . This makes sense physically since we would expect to require a dierent value for 1 , depending on which solution is chosen for 2 .

1.4. OUTLINE OF THE TEXT

25

Figure 1.17: The Epson Cartesian Robot. Photo courtesy of Epson.

Problem 4: Velocity Kinematics

In order to follow a contour at constant velocity, or at any prescribed velocity, we must know the relationship between the velocity of the tool and the joint velocities. In this case we can dierentiate Equations (1.1) and (1.2) to obtain 1 2 sin(1 + 2 )( 1 + 2 ) x = 1 sin 1 1 + 2 cos(1 + 2 )( 1 + 2 ). y = 1 cos 1 Using the vector notation x = equations as = x 1 sin 1 2 sin(1 + 2 ) 2 sin(1 + 2 ) 1 cos 1 + 2 cos(1 + 2 ) 2 cos(1 + 2 ) (1.11) x y and = 1 2 (1.10)

we may write these

. = J The matrix J dened by (1.11) is called the Jacobian of the manipulator and is a fundamental object to determine for any manipulator. In Chapter 5 we present a systematic procedure for deriving the Jacobian for any manipulator in the so-called cross-product form. The determination of the joint velocities from the end-eector velocities is conceptually simple since the velocity relationship is linear. Thus the joint

26

CHAPTER 1. INTRODUCTION

Figure 1.18: Workspace of the cartesian manipulator. velocities are found from the end-eector velocities via the inverse Jacobian = J 1 x where J 1 is given by J 1 = 1 1 2 s2 2 c1 +2 1 c1 2 c1 +2 2 s1 +2 1 s1 2 s1 +2 (1.13) (1.12)

in which c and s denote respectively cos and sin . The determinant, det J , of the Jacobian in (1.11) is 1 2 sin 2 . The Jacobian does not have an inverse, therefore, when 2 = 0 or , in which case the manipulator is said to be in a singular conguration, such as shown in Figure 1.29 for 2 = 0. The determination of such singular congurations is important for several reasons. At singular congurations there are innitesimal motions that are unachievable; that is, the manipulator end-eector cannot move in certain directions. In the above cases the end eector cannot move in the direction parallel to x2 , from a singular conguration. Singular congurations are also related to the non-uniqueness of solutions of the inverse kinematics. For example, for a given end-eector position, there are in general two possible solutions to the inverse kinematics. Note that the singular conguration separates these two solutions in the sense that the manipulator cannot go from one conguration to the other without passing through the singularity. For many applications it is important to plan manipulator motions in such a way that singular congurations are avoided.

1.4. OUTLINE OF THE TEXT

27

Figure 1.19: The ABB IRB940 Tricept Parallel Robot. Photo courtesy of ABB.

Problem 5: Vision
Cameras have become reliable and relatively inexpensive sensors in many robotic applications. Unlike joint sensors, which give information about the internal conguration of the robot, cameras can be used not only to measure the position of the robot but also to locate objects external to the robot in its workspace. In Chapter ?? we discuss the use of cameras to obtain position and orientation of objects.

Problem 6: Dynamics
A robot manipulator is basically a positioning device. To control the position we must know the dynamic properties of the manipulator in order to know how much force to exert on it to cause it to move: too little force and the manipulator is slow to react; too much force and the arm may crash into objects or oscillate about its desired position. Deriving the dynamic equations of motion for robots is not a simple task due to the large number of degrees of freedom and nonlinearities present in the system. In Chapter 6 we develop techniques based on Lagrangian dynamics for systematically deriving the equations of motion of such a system.

28

CHAPTER 1. INTRODUCTION Power supply Input device o or teach pendant

 / Computer o controller O  / Mechanical arm O 

Program storage or network

End-of-arm tooling

Figure 1.20: Components of a robotic system.

d d

Figure 1.21: Linear vs. rotational link motion. In addition to the rigid links, the complete description of robot dynamics includes the dynamics of the actuators that produce the forces and torques to drive the robot, and the dynamics of the drive trains that transmit the power from the actuators to the links. Thus, in Chapter 7 we also discuss actuator and drive train dynamics and their eects on the control problem.

Problem 7: Position Control

Control theory is used in Chapters 7 and 8 to design control algorithms for the execution of programmed tasks. The motion control problem consists of the Tracking and Disturbance Rejection Problem, which is the problem of determining the control inputs necessary to follow, or track, a desired trajectory that has been planned for the manipulator, while simultaneously rejecting disturbances due to unmodelled dynamic eects such as friction and noise. We detail the standard approaches to robot control based on frequency domain techniques. We also introduce the notion of feedforward control and the techniques of computed torque and inverse dynamics as a means for compensating the complex nonlinear interaction forces among the links of the manipulator. Robust control is introduced in Chapter 8 using the Second Method of Lyapunov. Chapters ?? and ?? provide some additional advanced techniques from nonlinear control theory that are useful

1.4. OUTLINE OF THE TEXT

29

Pitch

Roll

Yaw

Figure 1.22: Structure of a spherical wrist. for controlling high performance robots.

Problem 8: Force Control

Once the manipulator has reached location A. it must follow the contour S maintaining a constant force normal to the surface. Conceivably, knowing the location of the object and the shape of the contour, we could carry out this task using position control alone. This would be quite dicult to accomplish in practice, however. Since the manipulator itself possesses high rigidity, any errors in position due to uncertainty in the exact location of the surface or tool would give rise to extremely large forces at the end-eector that could damage the tool, the surface, or the robot. A better approach is to measure the forces of interaction directly and use a force control scheme to accomplish the task. In Chapter ?? we discuss force control and compliance and discuss the two most common approaches to force control, hybrid control and impedance control.

30

CHAPTER 1. INTRODUCTION

Figure 1.23: Angular Jaw Grippers.

Figure 1.24: Parallel Jaw Grippers.

Camera
A F

S Home

Figure 1.25: Two-link planar robot example.

1.4. OUTLINE OF THE TEXT

31

y2 y0 y1 2

x2

x1

x0

Figure 1.26: Coordinate frames for two-link planar robot.

elbow up

elbow down

Figure 1.27: Multiple inverse kinematic solutions.

32

CHAPTER 1. INTRODUCTION

c 1 1

Figure 1.28: Solving for the joint angles of a two-link planar arm.

y0

2 1 1 2 = 0 x0

Figure 1.29: A singular conguration.

1.5. PROBLEMS

33

1.5

Problems

1. What are the key features that distinguish robots from other forms of automation, such as CNC milling machines? 2. Briey dene each of the following terms: forward kinematics, inverse kinematics, trajectory planning, workspace, accuracy, repeatability, resolution, joint variable, spherical wrist, endeector. 3. What are the main ways to classify robots? 4. Make a list of robotics related magazines and journals carried by the university library. 5. From the list of references at the end of this chapter make a list of 20 robot applications. For each application discuss which type of manipulator would be best suited; which least suited. Justify your choices in each case. 6. List several applications for non-servo robots; for point-topoint robots, for continuous path robots. 7. List ve applications that a continuous path robot could do that a point-to-point robot could not do. 8. List ve applications where computer vision would be useful in robotics. 9. List ve applications where either tactile sensing or force feedback control would be useful in robotics. 10. Find out how many industrial robots are currently in operation in the United States. How many are in operation in Japan? What country ranks third in the number of industrial robots in use? 11. Suppose we could close every factory in the United States today and reopen then tomorrow fully automated with robots. What would be some of the economic and social consequences of such a development? 12. Suppose a law were passed banning all future use of industrial robots in the United States. What would be some of the economic and social consequences of such an act? 13. Discuss possible applications where redundant manipulators would be useful.

34

CHAPTER 1. INTRODUCTION

Figure 1.30: Diagram for Problem 1-15. 14. Referring to Figure 1.30 suppose that the tip of a single link travels a distance d between two points. A linear axis would travel the distance d while a rotational link would travel through an arclength as shown. Using the law of cosines show that the distance d is given by d = 2(1 cos()) which is of course less than . With 10-bit accuracy and = 1m, = 90 what is the resolution of the linear link? of the rotational link? 15. A single-link revolute arm is shown in Figure 1.30. If the length of the link is 50 cm and the arm travels 180? what is the control resolution obtained with an 8-bit encoder? 16. Repeat Problem 1.15 assuming that the 8-bit encoder is located on the motor shaft that is connected to the link through a 50:1 gear reduction. Assume perfect gears. 17. Why is accuracy generally less than repeatability? 18. How could manipulator accuracy be improved using direct endpoint sensing? What other diculties might direct endpoint sensing introduce into the control problem? 19. Derive Equation 1.5.9. 20. For the two-link manipulator of Figure 1.25 suppose 1 = 2 = 1. Find the coordinates of the tool when 1 =
6

and 2 = 2.

21. Find the joint angles 1 , 2 when the tool is located at coordinates 1 1 2, 2 .

1.5. PROBLEMS

35

1 = 1, 2 = 2, what is the 22. If the joint velocities are constant at velocity of the tool? What is the instantaneous tool velocity when 1 = 2 = 4? 23. Write a computer program to plot the joint angles as a function of time given the tool locations and velocities as a function of time in Cartesian coordinates. 24. Suppose we desire that the tool follow a straight line between the points (0,2) and (2,0) at constant speed s. Plot the time history of joint angles. 25. For the two-link planar manipulator of Figure 1.25 is it possible for there to be an innite number of solutions to the inverse kinematic equations? If so, explain how this can occur.

36

CHAPTER 1. INTRODUCTION